What are the advantages and disadvantages with filtering and sorting tables.
Are there specific situations where one is preferred over the other?
Can be determined by how inversely proportional the quantity to quality ratio is when comparing shear size of the dataset and individual characteristics of each item.
Useful in high quantities that also have
minimal deviation between common qualities that can
*differentiate* them from each other (such as color, weight, age).
Depending on the proportion of each, the two roles could just as easily switch. If you have a very large amount of seemingly unrelated data, an alphabetical sort obviously becomes unreasonable but can become very useful as a filter depending on the context of the data (filter all A's, all B's, all C - G, all except X, Y, Z)
lower/manageable quantities (often times as a result of a previous filter being applied) that have a high variation of generic qualities
*linking* one another (such as spelling similarities, numerical patterns or word size)
The inverse of this also applies: If your data is a
manageable amount but has different colors, sorting by color can be very helpful in navigating through as well.
At least...that's how I see it.
Filtering is good for "groups" of data like Departments, States, Categories, true/false (yes/no) types. Anything else that is more random like date and times, quantities, and numbers in general is better suited for a sort.
I only turn filtering on for fields that users explicitly request. If the user is "getting out of hand" (they never do that, right?) with filtering, I may even ask them how they would use the actual filter itself to see if it's worth talking them out of it.
I'd say filtering data is "more dangerous". Have you ever seen one of those crazy filter AutoSort columns in a huge Excel spreadsheet? I'd stay away from that type of design as much as possible. Make sure that the grid contains only the data that the user needs and use filters sparingly and only on the true groups of data.
Putting a sort on data that is highly variable (dates and numbers) is generally a best practice. Also, sorting, especially if you are paging your data, is pretty cheap for the most part so it doesn't really hurt to put it in.